Скачать с ютуб Machine Learning at Atlassian // Geoff Sims // Coffee Session#34 в хорошем качестве

Machine Learning at Atlassian // Geoff Sims // Coffee Session#34

Coffee Sessions #34 with Geoff Sims of Atlassian, Machine Learning at Atlassian. //Abstract As one of the world's most visible software companies, Atlassian's vast data and deep product suite pose an interesting MLOps challenge, and we're grateful to Geoff for taking us behind the curtain. //Bio Geoff is a Principal Data Scientist at Atlassian, the software company behind Jira, Confluence & Trello. He works with the product teams and focuses on delivering smarter in-product experiences and recommendations to our millions of active users by using machine learning at scale. Prior to this, he was in the Customer Support & Success division, leveraging a range of NLP techniques to automate and scale the support function. Prior to Atlassian, Geoff has applied data science methodologies across the retail, banking, media, and renewable energy industries. He began his foray into data science as a research astrophysicist, where he studied astronomy from the coldest & driest location on Earth: Antarctica. -------------- ✌️Connect With Us ✌️ ------------ Join our slack community: https://go.mlops.community/slack Follow us on Twitter: @mlopscommunity Sign up for the next meetup: https://go.mlops.community/register Catch all episodes, Feature Store, Machine Learning Monitoring and Blogs: https://mlops.community/ Connect with Demetrios on LinkedIn:   / dpbrinkm   Connect with Vishnu on LinkedIn:   / vrachakonda   Connect with Geoff on   / geoff-sims-0a37999b   Timestamps: [00:00] Introduction to Geoff Sims [01:20] Geoff's background [04:00] Evolution of ML Ecosystem in Atlassian [06:50] Figure out by necessity [07:51] "I think in general, as a newcomer to the field, you need to understand the maturity of the organization as to what is expected of you." [08:47] Machine Learning not priority number one and disconnected to MLOps [11:53] Atlassian being behind or advanced? [15:43] "Jira is more managing the running of the projects and whatnot." [16:38] Serious switch of Atlassian around machine learning [17:05] "It may be started as an experiment. To be honest, everyone is doing machine learning, we should probably do it as well. So, let's hire a bunch of people once that happens." [17:33] "Sometimes if you site enough times, it happens. If the right person says it enough times, it just becomes a fact. That's the way it is. That's what's happening." [17:47] What data org did it come from? [18:52] "Sometimes it's easy to be critical, but then given the state of that time, maybe it didn't make sense to have a fully-fledged data-role of 50-100 people." [20:00] Consolidation of the stack [20:49] "Even to this day, I don't think there has been a formal merge. Rather all those scattered people now have well-established teams they work with and the machine learning team also is a much more well-established team." [21:21] Tooling - blessing and curse [24:37] Tackling play out [25:55] "If you're storing everything, that's just the raw event stream. So you essentially need to recreate the process of producing features in an offline environment in order to simulate what these features would be in production." [27:35] "Ignoring the scale aspect was actually suitable for the online portion of the problem. It's not suitable for the offline portion because you couldn't use that to produce features for training a model." [29:38] Staying on the same page [30:48] Priority of needs [31:43] "The unfortunate side effect was that even when that was solved to the best of its ability, you have limited use and limited scale." [31:55] How did it evolve? [35:12] Where is Atlassian now? [36:06] "Atlassian built this formal data organization of our whole engineering arm. We are in a much better place." [40:21] "Architecturally, Tecton is very very similar (to ours), it was just way more mature." [41:17] What unleashed you to do now? [41:36] "The biggest thing is independence from a data science perspective. Less reliance and less dependence on an army of engineers to help deploy features and models." [44:25] Have you bought other tools? [45:43] "At any given time, there's something that's a bottleneck. Look where the bottleneck is, then fix it and move on to the next thing." [48:20] Atlassian bringing a model into production [50:01] "When we undertake whatever the project is, it's days or weeks to go to a prototype rather than months or quarters." [53:10] "Conceptually, you're struggling walking towards that place because that's the place you want to be. If that's your problem, that's good. That's the promised land." [54:45] "Using our own tools is paramount because we are customers as well. So we see and feel the pain which helps us identify the problems and understand them."